Re: Unambiguity proofs for CFG's

matt@waikato.ac.nz (Matt Melchert) writes:


>Last year I posted a query as to whether it was possible to determine if a
>grammar is unambiguous. The answer is that it is not, the proof being the
>Post Correspondence Problem. But further reflection has yielded the
>intuition that if "ambiguity" is equivalent to "can produce more than one
>parse tree for some input", then wouldn't it be enough to show that for a
>given grammar any input could still only produce one parse tree?


Yes, it would be enough to do that. It might even work for some grammars.
The problem would still remain undecidable.


The proof that I've seen (see, for example, Theorem 11.8.3 from Sudkamp's
excellent introductory text, "Languages and Machines") relies on the
definition that a grammar is ambiguous if it contains a string that can be
generated by two distinct leftmost derivations (which, as I understand it,
is the usual definition). This definition is equivalent to your definition
(there being two distinct parse trees), because any given parse tree has one
unique left derivation and that any left derivation has one unique parse
tree. (Exercise: prove this --- it's not hard)


Cheers,
Andrew Bromage
--

Return to the comp.compilers page.
Search the comp.compilers archives again.